Louis Rossmann on Sovereign Identity and Media Platforms

See Maki Kato Talking About Verifiable Credentials at VCF East and my comment:

This is cool. There are a lot of different ways to go forward from here too. One is to be able to handle credit for content creators. Everyone who uses social media, even if they are not themselves content creators, is doing useful work just by paying attention to content creators and integrating the different expressions of ideas that those content creators produce. A system which identifies content creators at a level above the particular platform they use is an essential component of any system that does this. The higher-level system that can be built on top of this identification produces a kind of web of influence, with currents of ideas of that circulate and are developed within it. These vector flows of influence are the basis for any proper system of recommendation which can provide a rational basis for recommending new sources of content, and one that is not so easily subverted by paid advertising and bot swarms which artificially amplify signals which otherwise would disappear into the noise. Obviously any such top-level system is itself a target for subversion, so it is important that what people understand are the ideas behind the operation of the system, and that those ideas are the real basis for knowledge, not just the concrete instances which some particular platform represents to them as "the truth".

Here's part 1:

Subscribe to Louis Rossmann.

Harpo Roeder discusses Neopass/Harbor. This is a really useful talk, he deals with a lot of systems at a very abstract level and you can get a much better idea of what the real problems are:

7:29 On Scuttlebutt (SSB) The problem of multiple devices. They ran into a problem because they based a user's timeline on a concrete series of representations of events recorded on some one particular device. Reintegrating timelines from multiple devices then became a complex problem. Imagine what happens when your phone loses connectivity with your desktop machine and you are away from your office and someone else sends messages from your desktop while you were out. Who sent those messages? How do you prove it wasn't you? What if it was you, and someone had stolen your phone and you didn't notice? You would need some attestation from someone who knows you to resolve the issue. See Maki Kato Talking About Verifiable Credentials at VCF East. What if, after losing your phone you had to set up new accounts on completely different media platforms and then establish those identities later and reintegrate them into your timeline, how are you going to do that?

13:42 Maybe RSS syndication isn't expressive enough? See this Wired piece: The Best RSS Feed Readers for Streamlining the Internet: "There are two parts to RSS: the RSS reader and the feeds from your favorite websites. RSS has been around awhile now, so there are a lot of very good RSS readers out there. Most of them feature built-in search and suggestions, so you don't have to go hunting for feeds yourself. You just might discover some cool new sites to read." Any syndication system needs to be able to handle "moods" adequately. Occasionally, I look up some pretty whacky stuff, but that doesn't mean I want all of the whacky stuff that's connected with it flooding my feed. I need to be able to partition influence in some natural way, connected with the internal structure of the flows. See Looks Like YouTube Are Doing Something Serious About Recommendations (but I don't think they are. The very concept of occasionally picking a global Google-doodle is evidence of that. How do you handle translations of the lyrics of songs, for example?). I had never heard of Elastic Search before. It sounds interesting: https://www.elastic.co/enterprise-search.

18:15 Recommendation Engines and trust. This is a good start, but ultimately I think trust is something that needs to be established organically.

19:20 The web of trust, ... My experience with that was low-quality! It was in 1999 when I first started working at the Computer Laboratory at Cambridge and Piete Brooks offered to sign my key, but he said that nobody there would sign for my iang@pobox.com email address. That was not on a system which was administered by either the Computer Laboratory or the Computing Service. I eventually lost that identity when I was in Bolivia in 2012 and there was nobody either prepared to or able to renew my subscription.

24:24 Using Web of Trust to model attestations in fact-checking exercises. He talks about using verifiable credentials a la Maki Kato Talking About Verifiable Credentials at VCF East.

See https://futo.org/ "Computers should belong to you, the people. We develop and fund technology to give them back."

Subscribe to Futo.

It's a great shame that there is such a gulf between the academic work on this subject and the technical work going on in the field. You would think that there in itself is a perfect opportunity to test the theory with practice, ... why doesn't this happen? See Robin Milner's Bigraphical reactive systems: basic theory (2001)

Peanut Butter Wolf - 02 - In Your Area (Ft Planet Asia)

Subscribe to @SethMethCS.